Multi-Scale Analysis Of The Relationship Between Commodity Prices And Macroeconomic Fundamentals In Sub-Saharan Africa

Abstract

This thesis explored the multiscale relationship between commodity prices (oil, gold, and cocoa) and macroeconomic fundamentals (inflation and exchange rate), focusing on predictability, systemic risk, information flow, and contagion. These issues are important to global investors seeking to reduce risk and maximise investment returns, as well as to policymakers who must craft appropriate policy responses even in the midst of financial turmoil. Multiscale relationships are assessed using a cross-quantilogram, complete ensemble empirical mode decomposition (CEEMDAN), transfer entropy, and Barunik and Krehlik (BK18) spillover index. These techniques are robust to weaknesses in methods like vector autoregressive (VAR) and generalised autoregressive conditional heteroscedasticity (GARCH) used in the prior literature. The results show that commodity prices have an inconsistent prediction pattern for inflation but are more successful at the extreme quantiles than the medium ones. Systemic risk from commodities reaches different countries at different times but is higher at the upper quantiles. Also, it was observed that the information flow between commodity prices and exchange rates is asymmetric and is both time- and frequency-dependent. Again, it was revealed that, while oil and cocoa are net transmitters of spillovers, commodity exporting countries' exchange rates dominate spillover propagation. The results also show that countries differ in their responses to spillover, which are both time- and frequency-varying. It was recommended that inflation and exchange rate policies be country-specific. Again, hedging and diversification strategies should be quantile dependent, frequency-dependent, and time-varying.